Download The Economics of Personalized Medicine and Genomics

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Oncogenomics wikipedia , lookup

Medical genetics wikipedia , lookup

Genomics wikipedia , lookup

Pathogenomics wikipedia , lookup

Pharmacogenomics wikipedia , lookup

Public health genomics wikipedia , lookup

Transcript
The Economics of Personalized
Medicine and Genomics
Robert D. Lieberthal, PhD
Assistant Professor,
Jefferson School of Population Health
Disclosures—JSPH Research Projects
• I have worked on projects at the Jefferson
School of Population Health funded by
Abbott
Laboratories (Abbott Molecular division)
Genomic Health (Oncotype DX®)
JSPH Academic Projects
Learning Objectives
1. Assess genomic approaches from the point of view
of a patient and a population
2. Critique current approaches to assessment of
personalized medicine
3. Evaluate the economic outcomes of genomic
medicine for different populations
Connections to Colloquium Sessions
• Carolyn Buck-Luce, MBA
Global Pharmaceutical Sector Leader, Ernst & Young, New
York, NY
The Importance of Innovation for Life Sciences
Tuesday, 2/28, at 11am
• Chris McFadden
Managing Director, Health Evolution Partners, New York, NY
Closing Keynote: Healthcare Investment Trends
Wednesday, 2/29, at 8:15am
Genetics is a major determinant of
population health outcomes
Health Care
Environment
Genetics
Social
Behavior
• Source: Population Health, Ch. 10
• Behavior is number 1 at 40%
• Genetics is high at 30%
Genetics versus Genomics
• Genetics
“Study of genes and their roles in inheritance”
Genetic diseases: Cystic fibrosis, Huntington's disease, and
phenylketonuria (PKU)
• Genomics
“Describes the study of all of a person's genes (the genome)”
Complex diseases: heart disease, asthma, diabetes, and cancer
Combination of genetic and environmental factors
“Genomics is offering new possibilities for therapies and treatments for
some complex diseases, as well as new diagnostic methods.”
Source: National Human Genome Research Institute, National Institutes of Health; genome.gov
Diagnostics as the first step
toward Genomic Medicine
• We have mapped the whole human genome
• Reasonable first step: relate that map to known
•
illnesses
Genomic diagnostics
Predictive genomic tests—Oncotype DX® (Genomic
Health)
Genomic therapies—Vysis ALK FISH test (Abbott
Molecular) and crizotinib (Pfizer)
Population health purposes
for Genomic Diagnostics
• Research
This
is where we are now
• Improvement
Future
medicine—combine data from success of
drugs with multiple genomic tests
Refine treatment for subpopulations
• Accountability
Definitely
not yet
Examples: herceptin—HER2 breast cancer
Genomic tests sort people out
• Think about oncology drugs
Traditional chemotherapy
Novel, targeted, molecular therapies
• How much would a test be worth that separated responders
from non-responders?
Direct valuable healthcare resources towards those most likely to
benefit
• Value of test depends on population level variables
What percent of people are expected to succeed and fail?
Test validity is crucial and dependent on the scenario
Genomic diagnostics can help choose between
existing options
• Genomics can show commonalities within groups
Genetic background
Gene-environment interaction
Special mutations
• Genomics is about classification
Which similarities matter clinically—go beyond disease diagnosis
What worked in the last patient that will work with this patient?
• Diagnostic genomics is an emerging field
Familiar challenge—making bench science into clinical therapy
New challenge—find which common markers that are already known
are valuable
Examples of genomic diagnostics
Oncotype DX®
• Genomic Health
• Risk score for breast and colon cancer
• Stage 1 or 2 estrogen receptor-positive, lymph node-negative
breast cancer
• Combines assay of 21 genes with “scoring” algorithm
• Generates recurrence score between 0-100
• Behavior change potential
Determine recurrence probability
Evaluate likely benefit from chemotherapy
Source: “The Economics of Genetic Testing for Women with Breast Cancer” (2012). Working paper, Jefferson
School of Population Health.
Examples of genomic diagnostics—Vysis ALK
Break Apart FISH Probe Kit
•
•
•
•
•
•
•
Abbott Molecular
Non-small cell lung cancer
FISH test—genomic test to map genes
ALK—a gene implicated in many types of cancer
Predicts response to specific drug—crizotinib
FDA approved, marketed as Xalkori® by Pfizer
Behavior change potential
Evaluate benefit from a targeted molecular therapy that
benefits a minority of patients
Source: FDA News Release, “FDA approves Xalkori with companion diagnostic for a type of late-stage lung cancer”
www.fda.gov/NewsEvents/Newsroom/PressAnnouncements/ucm269856.htm
Targeted therapies aren’t for everyone
• Successful genomic therapies often look great
From a clinical point of view
Ignoring the fact that we have to figure out how to assign
treatment
• Successes and failures will tend to fall in groups
Can we identify those groups ahead of time?
If not, how rigorous are our post-hoc analyses?
• The main point
What looks great to a single patient doesn’t always look
great on a population level
Personalized Medicine
means recognizing and using options
• Conventional therapies
Has the richest evidence base
• Novel therapies
Therapies that are often costly and/or require special training
• Alternative/complementary therapies
• Watchful waiting and/or palliative care
Not everything is curable (or even treatable)
A legitimate therapy for many conditions
Doctors may have the most work to support patients in this option
The outside option in most cases
New treatments complicate
Personalized Medicine
• Complications for study design
More options to evaluate
More subgroups to consider
• Complications for creating an evidence base
Evidence-based medicine is based on matching condition and
treatments
Personalization makes it harder for clinicians to look at the right “cell”
in the evidence database
• Complications for economic analysis
Economics depends on the marginal choice
It’s harder to identify the “marginal choice” for each patient
Where will policymakers get their data?
Regulatory “Wild West”
• FDA does not regulate laboratory
developed tests
• Direct-to-consumer tests are
available
Navigenics, 23andMe, and
Decode Genetics
These companies have called for
greater regulation
• Potential issues
Quality control
Reimbursement
Lack of development of FDA and
other expertise
• Can personalized medicine
succeed without regulation?
Billing for genomic tests is complicated
• No stand-alone CPT codes
ICD 10 may address this issue
• Example: ResponseDX CPT codes
Lung ® : 88323, 88381, 88313, 83907, 83891, 83902, 83898, 83896,
83912, 83900, 83901
Colon ® : 88323, 88381, 88313, 83907, 83891, 83902, 83898, 83896,
83912
Gastric ® : 88323, 88381, 88313, 83907, 83891, 83902, 83898, 83896,
83912
Melanoma ® : 88323, 88381, 88313, 83907, 83891, 83898, 83896,
83912
• Can personalized medicine succeed if we can’t bill for it?
Contemporaneous approval of
therapy and diagnostic
• FDA is pushing for contemporaneous approval of therapy and diagnostic
• Two oncology approvals in 2011
Vemurafenib (melanoma) and cobas 4800 BRAF V600 Mutation Test
Crizotinib (lung cancer) and Vysis ALK Break Apart FISH Probe Kit
• Personalized medicine issues
Do those who “fail” the test get the drug?
• Economic issues
Evaluate the two simultaneously
Impossible to separate the comparative effectiveness of the test from the drug
under current regulations
Back to the question: how do we value the test?
Answer to personalized medicine issue dictates the setup of economic
evaluation
An economic perspective is needed
• The marginal value of extra information
Not how good is genomics
How much more good does it provide?
Value could be financial, clinical or humanistic
• Getting homogenous subpopulations is the point of
personalized medicine
Not just on observable characteristics
Estimate what variation is unobserved
Maximize the value of all information
That describes the techniques for observational studies in economics
research
Economics is about Value
Low monetary cost
High monetary cost
Low non-monetary
costs
Potentially high
value treatment
High financial cost
High non-monetary
costs
High
clinical/humanistic
cost
Potentially low value
treatment
• All cells require assessment of benefits
• All cells require additional comparative effectiveness analysis
Genomics doesn’t have to be fancy and expensive
• Clinicians already collect a lot of data—i.e. oncology
Sex and gender
Race and ethnicity
Age
Cancer stage
Cancer histology
• Usually cheaper to collect data by asking people than by
•
•
running assays
The promise of genomics is getting beyond the plainly
observable
Genomics imposes many costs to get additional data
Vysis ALK FISH test requires a tumor sample
Value of genomics depends on
modality of practice of medicine
• Whose outcomes are we maximizing?
The patient
The statistical patient
Population health
• Do we personalize as much as we can?
We take into account as much data and experience as possible
Then we still have lots of partially informed choices—possibly with
equally proven outcomes
How should payers choose between me-too drugs and me-too
therapies?
• The answers could make genomics more or less valuable
Minimizing societal costs
• Economic perspective
• The value per average person
• This will make a test worth less or zero if there are too few or
TOO MANY successes for our hypothetical drug
• Think of Vysis ALK FISH test
Too few predicted responders—the test identifies few new treatments
Too many predicted responders—we are harming people by charging
for the test when we should just give everyone crizotinib!
Despite the heavy cost of inappropriate treatment for those nonresponders
U.S. perspective on who bears the cost
• Treatment failures (or their insurers) pay a cost and get no
benefit
• Treatment successes (or their insurers) pay a cost and get a
benefit
• Different rule
If Oncotype DX® test result changes treatment path, it was
more valuable for that patient than average
You could charge successes once the test comes back, or
charge everyone upfront and rebate the failures (VBID/risk
sharing/other)
Universal health care systems perspective
on who bears the cost
• Societal perspective
• Individuals accept that the payer may pay for treatments that
don’t benefit them
• The whole point is that we can easily identify the
beneficiaries—more so than in many other types of medicine
• Equity issue—are those who don’t benefit from the test left
out?
Some people may be harmed by new technology
• Common problem in U.S., universal systems
• Some may pay for a test that doesn’t directly benefit
them
• Some who would have benefited from the old drug
get the new drug
• Economics: how could we balance these harms?
Place of health policy is to raise these issues and make
sure no one is harmed too much
• Genomics in personalized medicine
It’s a science
Health policy can’t change science
• The rationale and implications of the economic approach
Lots of studies involve modeling
We want to cut down to binary choices through the marginal approach
• Policy approach
Balance outcomes for populations and existing patients
Consider intended and unintended consequences
Decide how much weight to put on observational studies
Benefit the population
Summarizing the economic approach to
Personalized Medicine
• Economic evaluation depends on
The practice of medicine
Regulator behavior
How medical science evolves
Population versus patient focus
• Test gives valuable information on whether to proceed with
investment (expensive new treatment)
• Price discrimination may help access but hurt equity
• VBID/risk sharing arrangements—“no outcome, no income”!
Learning Objectives—Review
1. Genomic approaches
a)
b)
Patient wants personalized treatment
Population dictates how much variation we can observe
2. Assessment of personalized medicine
a)
b)
c)
Regulators have some expertise but limited say
Payers can’t find some treatments in their claims data
Some may resist evidence-based medicine, and EBM may not be
informed for every group
3. Evaluate the economic outcomes
a)
b)
c)
Some RCT evidence
More observational studies/modeling
Policy on how technology is deployed dictates how it should be studied